YouTube announced Wednesday that it will automatically apply AI labels to videos when its internal systems detect significant photorealistic AI use, even if a creator has not disclosed that AI was involved in making the content.
The platform will apply labels when its systems detect significant photorealistic AI use, even if creators don't disclose it
YouTube announced Wednesday that it will automatically apply AI labels to videos when its internal systems detect significant photorealistic AI use, even if a creator has not disclosed that AI was involved in making the content.
Existing rules obligating creators to flag realistic AI content have not been dropped, with the automated detection functioning as an additional enforcement layer. When a creator skips the disclosure step entirely, YouTube's detection systems can step in and attach a label on their behalf. For any creator who feels the system got it wrong, a correction can be submitted through YouTube Studio by adjusting the video's disclosure settings.
In some cases, however, labels will be permanent. Permanent exceptions exist, however: content produced using YouTube-native tools such as Veo and Dream Screen will retain its label regardless, as will any video carrying C2PA metadata that identifies it as entirely AI-generated.
YouTube is also changing where AI labels appear. On long-form uploads, the label will be repositioned so it sits between the video player and the description text. In the case of YouTube Shorts, the label will be embedded directly over the video content. Previously, labels appeared in the expanded description for most content. Videos that fall into less realistic categories — such as animated or minimally altered content — will have their disclosures tucked into the expanded description as before.
Rene Ritchie, who serves as YouTube's head of editorial and creator liaison, addressed the updates in a dedicated video, stressing that the labels carry no algorithmic or financial consequences for creators: "do not affect how our videos are recommended or whether they can earn money. This is purely about giving viewers the right information at the right time."
YouTube said the changes are designed to balance transparency with creator control. The practice of labeling AI content dates to 2024, when YouTube first introduced a Creator Studio feature that put the onus on creators to flag material realistic enough to be confused with genuine depictions of people, places, or events, according to TechCrunch.
YouTube said the new automatic detection is rolling out in May 2026. Under the old arrangement, only videos covering sensitive areas such as health or news received the more visible treatment, with everything else relegated to the description; the updated system extends prominent labeling to all photorealistic or substantially AI-generated material, according to Variety.
Join 500,000+ readers who start their day with Quartz.
By subscribing, you agree to our Terms of Service and Privacy Policy.